15 Jan 2012

What is Bounce Rate and Why should we worry about Bounces?

As per the "Content Characterization" section of Web Analytics Definitions by Web Analytics Association (www.webanalyticsassociation.org):

TERM: Bounce Rate
Type: Ratio
Universe: Aggregate, Segmented
Definition/Calculation: Single page view visits divided by entry pages.
Comments: If bounce rate is being calculated for a specific page, then it is the number of times that page was a single page view visit divided by the number of times that page was an entry. If bounce rate is calculated for a group of pages, then it is the number of times pages in that group was a single page view visit divided by the number of times pages in that group were entry pages. A site-wide bounce rate represents the percentage of total visits that were single page view visits.

That's the standard definition. But, it would have been better if we could measure "the percentage of website visitors who stay on the site for a small amount of time (usually five seconds or less)". This less than 5-seconds stay on the site is difficult to measure. This is because all Web-Analytics Software calculate the Time-on-Site by tracking the time spent on previous page when Visitor clicks on the next page. If the Visitor did not click on next page (bounce), how do we know the time spent on single page visits? Alas, if we could measure it somehow !

Anyway, bounce rate is a interesting way to measure the quality of traffic coming to a website. In short bounce rate measures the percentage of people who came to your website and left "instantly". Bounce rate measures quality of traffic you are acquiring, and if it is the right traffic then it helps you understand that where and how your website is failing your website visitors.

What is a Bounce?
Now, lets try to understand that how the Bounces happen:
  1. Any click on the page that directs a user to an external website or your sub-domain (yes, sub-domains are counted into your bounce rates).
  2. Pushing the back button and going back to the source.
  3. Closing the browser tab or the entire browser.
  4. Typing a new URL from that page and leaving.
  5. Timeout session, which is more than 30 minutes on a single page (in my opinion this one is little misleading factor, it can mean both that the user is inactive or that the page is so interesting that it engaged the user beyond 30 minutes - for example, a page having interesting large Video - this must be studied while analyzing bounce rate for the site in question).
All these are taken into account when calculating your bounce rate compared to the total number of visits to a single page.

By the way I would like to emphasize here that we should not get confused between Bounce-Rate and Exit-Rate. You may like to read Web Analytics Bounce-Rate and Exit-Rate.

What is the industry standard for bounce rate?

Short answer is that there is no industry standard. There are a lot of factors that influence the bounce rate, so you really can’t compare bounce rates of one site (or page) to another. The best way to know if you are doing better or worse is to set your own baseline and compare your performance over time.

Here are some of the numbers that were listed by Steve Jackson (Co-Chair Nordic Branch, Web Analytics Association), based on his experience with various sites.
Source: http://tech.groups.yahoo.com/group/webanalytics/message/6116

What are the factors that affect the bounce rate?

Below are some of the factors that determine the bounce rates. You can use this as a checklist to diagnose a high bounce rate issue. (Source: http://webanalysis.blogspot.com/2007/07/bounce-rate-demystified.html)

1. Source of your traffic – Each source results in a different bounce rate. When setting your baseline create overall baseline and baselines for each traffic source e.g. display advertising, organic traffic. Segment the Bounce Rate with Traffic Sources and then Analyze.

Do you need to revisit relationships with sites that are not sending you high quality traffic? What is the call to action that is causing people to come to your site and bounce? Are your email, affiliate, other marketing campaigns yielding low bounce rates?

2. Search engine ranking of the page – A page which ranks higher on irrelevant keyword will get a higher bounce rate. Measure the bounce rate of your search keyword.

3. Type of Audience – If you are advertising and reaching the wrong audience you will see higher bounce rate. Bounce rate will tell you if you need to better target your ads.

4. Landing Page Design – Landing page design affects the bounce rate. I suggest A/B testing to improve after you have set your baseline. No matter how low you go there is always an opportunity for improvement unless you somehow achieved 0% bounce rate.

5. Ad and Landing Page Messages – If the messages on your banner or search ads are not aligned with the messages on the landing page then the chances are you will have one of those 50% + bounce rates. Make sure messages are aligned and give visitors a clear call to action. Many a times marketers send users to a generic page instead of an appropriate landing page. This can (and will) result in higher bounce rates. Again A/B or multivariate testing should be used to reduce the bounce rate.

6. Emails and Newsletters – Subject lines, to and from, links, banners, the layout of email and the landing pages all work in tandem. They can either result in a great user experience and hence lower bounce rate or can result in a disaster. Do testing to reduce bounce rate.

7. Load time of your page(s) – A longer load time can result in visitor bailing out of the site causing higher bounce rates. Conversely, users can hit the refresh button, thinking there was a problem with the page load. This will incorrectly reduce bounce rate.

8. Links to external sites – A page that has links to external sites (or sub domains or pages that are not tracked in the same data warehouse) will show higher bounce rates.

9. Purpose of the page – Some pages’ purpose is to drive users inside the site while other pages provide the information that user is looking for. A page that provides the end result can show higher bounce rate. One example is the Branch-Offices page on my company’s web site, I have this page bookmarked. Whenever I need a particular branch's phone number, I go to my favorites, pull this page, get the number and leave.

10. Other factors - Pop-up ads, pop-up survey requests, music, streaming video, all can have an adverse effect on bounce rates if users become annoyed.

Worry about Bounce-Back rater than Bounce-Rate

First I want to make a couple of things clear. It’s highly unlikely that search engines use bounce rate directly when scoring or ranking webpages. Nor is a high bounce rate a definite signal of low quality or a failure to meet visitor expectations or needs.

Over the past few years, there has been a lot of debate (and confusion) about whether bounce rate is a signal that the search engines use to determine quality content. Matt Cutts has probably been asked this question ten thousand times, and has referenced bounce rate when answering some questions about ranking factors. Matt Cutts has said that bounce rate is spammable and a noisy signal. I also think that bounce rate is a poor signal for Google's use.

Just think about all of online calculators, Time Zone Converters or a Video article. They can have extremely high bounce rates since they can satisfy the user with only 1 page view needed. If your pages are satisfying the user I would not worry about bounce rate.

I have seen many instances of pages with high bounce rates (in Google Analytics) that still ranked well in Search. The content I’m referring to would clearly be identified as high quality, unique, and valuable, but had high bounce rates. Many of the pages I am referring to ranked in the top three to five listings on page one of Google, Bing, and Yahoo.

Duane Forrester, Sr. Product Manager at Bing as part of a post (How To Build Quality Content), explains that the engines can monitor “dwell time”, or the time a person remains on your page before clicking back to the search results. If visitors are clicking through the search listings to your site, and then clicking back to Bing quickly (in just a few seconds), that can be a negative signal to Bing.

Google internally refers to a Visitor that doesn't bounce back as a "long click". A long click is a click that leaves Google and doesn't come back for a long time: until that person wants to use Google for an unrelated search. The user doesn't refine their Keyword, nor does the user use the back button and click another result in the SERPs instead.

If Google is trying to use Bounce-Back as a ranking signal, they will have to deal with some complications -- A lot of people will rapidly open a series of tabbed results by repeatedly jumping back to the SERPs page. They won't even look at any of the pages until they finish opening a series of tabs. This makes it appear that rapid bounce-backs are occuring when they aren't.

In my opinion Google is clearly able to handle this multi-tab result opening. If I do a search, click on the result and then use the back button, I get a little notice in the SERPs to "block all <site> results". If I open many sites in tabs, I don't get this notice. So Google is able to penalize the Bounce-Backs. I believe that Google started using this as site wide signal as a big part of panda: "Bounce-Back from so many search terms sent to your site is high, the site (or a section thereof) must not have high quality content at all."

My recommendation for Bounce-Backs: If your traffic is coming to pages with high bounce rate with irrelevant keywords, try to adjust the content so that it should not rank for those irrelevant keyword. In case Bouncing Traffic is coming to a page with Keywords with different intent, give links to relevant content even if it is outside your site. This will reduce Bounce-Back, and the Pages / Site will ranks high.

Currently, there is no methodology available for the Webmasters to track Bounce-Back. Bounce Rate is easy to measure and can be a great proxy for Bounce-Back or Long-Click on a site with deep content.

Hopefully, analytics packages will evolve to let us see Dwell Time or Bounce-Back or Long-Click. In the meantime any measure that can be employed to improve engagement and increase the time visitors spend interacting with our content is essential.

No comments:

Post a Comment